Survey of Web Database Clustering Techniques

Nowadays databases have become web accessible through HTML form-based search interfaces The data units returned are encoded into the result pages dynamically for human browsing. For the encoded data units to be machine processable, which is essential for many applications such as deep web data collection and Internet comparison shopping, they need to be extracted out and assigned meaningful labels. Deep Web contents are accessed by queries submitted to Web databases and the returned data records are enwrapped in dynamically generated Web pages. Extracting structured data from deep Web pages is a challenging problem due to the underlying intricate structures of such pages. Author proposes some automated annotating approaches to improving search method on basis similarities in data units. They can form groups of data units and label them with semantic names and can store result for future analysis also.